On Thu, 2003-11-06 at 23:31, [EMAIL PROTECTED] wrote: > I guess in this case the hashes aren't really hashes anymore, they're more > like arrays. There wouldn't be any bucket under hash unless the key was > identical for two values. The B-tree entry becomes x div 4 (or an > appropriate number). This might allow quicker equality tests without > compromising less than and greater than tests. Is this the kind of > structure you're endorsing?
Sorry; div is right... It's weird- I had a dream about this last night... It may be that each of those tree roots should either be additional B-tree structures or a mirror of the topmost structure with a different function. Certainly: they should never be a "linked list" Further: it may be easier to think of the function as simply producing an index into the B-tree... --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
